Re: Point Cloud Info Needed (fwd) oops

From: Justin R. Kidder (
Date: Tue Jan 13 1998 - 19:35:00 EET

That should be (# of triangles sharing a vertex) times as many
entries at each vertex. My bad.

-Justin :)

Justin Kidder, Research Asst. | Automation and Robotics Laboratory | University of Pittsburgh
               Home page:
       "Nothing is fool-proof to a sufficiently talented fool"

It's also important to note that if you are choosing the vertex locations,
some data reduction should probably be done--coming from STL, you have 3
entries at each vertex. A good point cloud would not have these
redundancies. Assuming that your STL file has no errors, it would not be
a problem to find the repeats, but if you have accuracy errors in your
vertex locations, you may need a more robust method of detecting the


On Tue, 13 Jan 1998, Delft Spline Systems wrote:

> >When creating a point cloud file from a STL file, are the points the center of
> >the triangles or the three coordinates ?
> >
> Basically you are of course absolutely free to create any type of point
> cloud file from your STL: it may even contain many points on the
> surface of each separate triangle to have a sufficiently dense point cloud.
> In case the density is no problem, than the best apporoximation
> of the original CAD model is given by the three vertices of every
> triangle (of course three coordinates per vertex). These points are
> on the original CAD-surface, while the flat triangle inbetween is just
> an approximation.

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:44:42 EEST